Phil Westcott: From 'Batch Job' to 'Continuous Intelligence' in Deal Origination
Summary In this episode of the AI Pathfinder for Private Equity podcast, Steve Budd interviews Phil Westcott, CEO and co-founder of Filament Syfter. They discuss the transformative role of AI in private equity, focusing on how Filament Syfter helps firms build proprietary market intelligence to enhance deal origination. Phil shares insights on the inefficiencies in traditional processes, the importance of integrating data, and the competitive edge that AI can provide. The conversation also to...
Summary
In this episode of the AI Pathfinder for Private Equity podcast, Steve Budd interviews Phil Westcott, CEO and co-founder of Filament Syfter. They discuss the transformative role of AI in private equity, focusing on how Filament Syfter helps firms build proprietary market intelligence to enhance deal origination. Phil shares insights on the inefficiencies in traditional processes, the importance of integrating data, and the competitive edge that AI can provide. The conversation also touches on the cultural shifts within private equity firms as they adapt to new technologies and the strategic steps they can take to leverage AI effectively.
Takeaways
- The traditional deal origination process is often inefficient
- Private equity firms need to integrate their data for better insights
- AI can help firms identify and track the best fit deals
- Building a proprietary engine is crucial for competitive advantage
- Human insight remains essential in the AI-driven process
- Cultural shifts are necessary for successful AI adoption in private equity
Chapters
00:00 Introduction to AI in Private Equity
02:33 The Journey of Filament Sifter
05:22 Identifying Opportunities in Private Markets
07:51 Transforming Deal Origination Processes
10:31 Building Proprietary Market Intelligence
13:09 Leveraging AI for Competitive Advantage
15:51 Measuring ROI in Deal Origination
18:31 Cultural Shifts in Private Equity
20:40 Next Steps for Private Equity Firms
23:17 Future of Data and AI in Private Equity
Links
Phil's LinkedIn: https://www.linkedin.com/in/philwestcott/
Filament Syfter: https://filament.ai/
AI Pathfinder Private Equity AI Strategy Meet-ups:
https://www.aipathfinder.co/private-equity-ai-events
Have a question or would like to suggest improvements, please contact steve@aipathfinder.co
Well, hello and welcome to the AI Pathfinder for Private Equity podcast. If you've been following this series, you'll know I'm on a mission to help private equity firms move beyond AI hype and start applying it with confidence. One approach I've found really valuable is talking directly to founders, especially those building narrowly focused solutions for problems that private equity professionals deal with every day. These are often people with deep domain expertise, solving for a clearly defined job to be done. And they bring a unique perspective on how AI can deliver real, measurable value. And today's guest is exactly that. Phil Westcott is a CEO and co-founder of Filament SIFTA, an AI-driven platform that helps PE firms create and continuously monitor their own proprietary market intelligence to supercharge dual origination. Phil, welcome to the podcast.
SPEAKER_01Thanks, Steve. Thanks for having me. Excited about the discussion today.
SPEAKER_00Absolutely, absolutely. Well, look, let's kick things off straight away. Can you tell us a bit about yourself and the journey to a private equity workflow solution with a filament sifter?
SPEAKER_01Yeah, well, I've I've had a 25-year career in tech, in building the business cases, building the use cases around cutting-edge technology. I got into AI back in my IBM Watson days about 15 years ago. Watson, as you recall, was the first of the big AI platforms to come to market. And those early days were about, as I say, finding the use cases, the business cases, and getting those first scores on the board.
SPEAKER_00Well, not much has changed there, has it? You know, I think people are still grappling with where best to deploy AI. But you've found you've found a particular niche, haven't you?
SPEAKER_01Yeah, so we we founded Filament back in 2016. At that stage, our initial business model was innovating and consultancy-based. We would build prototypes for clients. We did a lot of work in public markets initially, building market signal platforms that would enable public market investors to detect why there had been a price change or why a stock had had a reaction in the market. As was the case from the very beginning of AI, the concept is in order to get alpha as an organization, you have to bring your data in-house. And that becomes the fuel for your AI roadmap.
SPEAKER_00Okay, and well look, let if we just drill down a little bit into the area that you've you've narrowed down on. What did you find then? Well, why that area? You know, what what's broken there you felt that could be um you know that's the same thing.
SPEAKER_01Yeah, so we got into private markets with a private equity firm back in late 2017. Uh, and the exam question there was how can we better sift through the huge numbers of opportunities in the market to focus our efforts on the best fit deals for their investment strategy? So that concept remains uh today our value prop. But since 2017, obviously, we've we've built up a huge body of work and experience in working with private market data and delivering that value to private equity firms alongside a few corporate finance houses that work closely with private equity in finding those best fit deals in the private markets and feeding their their deal engine and ultimately deploying capital or winning mandates.
SPEAKER_00Is it the case that that something was broken there, or is it just the opportunity to far improve the process?
SPEAKER_01Yeah, I I wouldn't say broken per se, but clearly inefficient and and more to the point, particularly in the last year or two, the forward-looking private equity firms have taken steps to build that in-house capability, that proprietary deal engine. Why is it accelerated of late? I think there's two two reasons. One is the hugely competitive market in private equity, the competing for LP funding, having an edge, uh, getting to the best fit deals before the competition, but also the advancements in technology and making sure that you as a fund are ahead of the game on leveraging this incredible new suite of Gen AI tech.
SPEAKER_00And it's it's not necessarily, well, it's about improving the process, but just reading one of your blog articles about the potentially array of third-party data sources that come in. There's a time and effort saving here, isn't there, potentially?
SPEAKER_01It is absolutely. I mean, I would strongly advocate that all private equity firms are going to need to have their proprietary market data set, i.e., their top of funnel captured in-house, and moving origination from a batch process to a continuous monitoring of the market, codifying their strategy and finding those best deals in the market. In order to do that, there is a piece of work, which is the complex challenge we've been iterating on and improving over those eight years, which is the data engineering and the integration piece. So, yes, it's about peering into the fog of private markets. We all know it's an incomplete data set, but the detective work of piecing together the different uh structured and unstructured data sets, leveraging some of the really powerful uh private market data sets you can buy on the market today, really good tools, many of which we partner with, which focus on bringing a particular insight into a corner of the market. Now, most of those kind of solutions are SaaS, but they all provide their data via API or via other provisioning, Snowflake, and other means. The challenge for the private equity firm is how do we bring that in-house and how do we codify our strategy so we can turn origination and deal making into an automated approach?
SPEAKER_00Would you say then that traditionally it was an inconsistent um process, potentially rather than something that you're making quite structured and and well based on facts and can run the same over and over again? Um, you know, I'm just wondering if if if that's if previously is that there's inconsistencies there. And and and I suppose I suppose my uh uh ultimate question is is it that you know the real sort of burning platform is they they're missing opportunities versus this is a case of of tidying up a process?
SPEAKER_01Yeah, I think it's a bit of both. So so part of it is that that pain of missing a deal that would have been perfect for your fund. And it's happened to every GP in the market at some at some stage and some more often than others. Ultimately, the old way of work was was human grunt work, scouring the different platforms, looking for good fit deals, throwing origination resource at the challenge to make sure you didn't miss a perfect opportunity. Even firms that's origination style has been mainly leveraging the bankers, bringing them deals, in order to win those deals, you need to have developed a good relationship with the founder and the CEO, sometimes two, three years ahead of the transaction. So doing it in the old way requires a bit of luck, a bit of human skill, a lot of grunt work. Um, but it's obviously now that there is huge opportunities to help make that whole process of finding and then monitoring the best deals for your for your firm in into a more systematic approach. As I said earlier, turning it from a batch job of human effort into a continuous operation of tracking the market for your deals and tracking the known opportunities for the right moment to reach out.
SPEAKER_00That lead time can be quite extensive then. So you were talking three years potentially.
SPEAKER_01Yeah, I mean, some deals from when you first meet the founder could take that long. Some obviously, if you're if you're lucky, you just find them at the right moment, can be a lot shorter. But either way, having found the opportunity, a huge piece of winning the business, winning the deal is tracking them and nurturing that relationship with the founder and the CEO. So having those nuggets to reach out and congratulate them, especially in this kind of Chat GPT outreach world where you know it's uh you're spammed continuously, uh, as we all are now, having that personalized outreach where you actually feel that investor or that corporate finance firm, that banker knows my business, understands me, and and would be a good partner when I do transact.
SPEAKER_00Yeah, I can see that. Well, let's go um see if we could get a bit under the hood then in terms of filament sift uh um could you walk us through, you know, I suppose step by step in the time that we've got here, how filament works?
SPEAKER_01Yeah, so essentially we are the that backbone to private equity and corporate finance building their proprietary top of funnel database. So it's their their deal engine. We are always white labeled, we're always deployed into their tech stack, but we're enabling them to build their proprietary asset. So once a PE firm has decided to go on that on that journey, they've recognized the operational and strategic imperative to do so. We are by far and away the most effective mechanism for them to get to that outcome faster with less risk and bring the benefits of our eight years of experience of building these deal engines in private markets.
SPEAKER_00Uh, just going back to the the um the proprietary piece, how how long can that take? For some, I assume it's just in you know, in their their heads, or it's not necessarily written down. I mean, it it that I can imagine is a really sort of um uh important process for them to go through anyway, so to codify exactly what you know what it is that they uh they're they're looking for.
SPEAKER_01Yeah, so back back in 2016, we we built from the ground up, uh so we know how long it takes, and we've we've obviously spoken to a number of our clients that originally tried an in-house build approach. You're talking uh 15 to 18 months to get operational, you're gonna need data engineers, data scientists, product managers, program managers. It is an in-house build is a expensive and time-consuming process, and that is really my key takeaway from this podcast is you don't need to build it on your own. Absolutely, you need a proprietary engine, but tools like Filament SIFTA are exactly built to enable that process. We've we've effectively productized the data engineering element of that architecture so that our clients can focus their energies on the secret source, as you say, codifying their strategy into algorithms and models. We facilitate it, but the business logic and the IP is very much the client. So this is their platform, they will always name it. Um, within in-house, within the colleagues, it would be considered to be the in-house engine, it's just powered by SIFTA in the background.
SPEAKER_00And is the yeah, obviously the the network that I have is it's obviously AI flavoured, AI Pathfinder. How much is AI sort of behind the scenes enabling what you do as well as making it very accessible?
SPEAKER_01Yeah, so the the big point here is that we are helping the PE firm to build their AI. Now, as you'd expect, given our heritage from our IBM Watson folks and our large data engineering and AI team uh based down in Southampton, we have solved the problems of data integration in private markets using our AI. But the the USP is the fact that with this infrastructure, the PE firm is actually building their asset, their proprietary data they own internally. They're enriching it with every interaction, and that's allowing them to train their own proprietary AI and gen AI tools. Back on the integration point. So how it works is we've we've integrated with 32 different private market data sources across news, transactions, financials, firmographics, people to enable that to get into an internal database. Think of it like a snowflake that has been verticalized for private markets. So you've got that internal, uh, you've got that internal engine, which is assimilating all the structured and unstructured data from the market, but then also the internal integrations. So it is feeding the CRM, it is integrated with the document management system, frequently SharePoint or ignite, and allowing them to enrich all their internal systems with live market data.
SPEAKER_00Okay, and I can see that, well, with the way you're describing it now, this could go far beyond dual sourcing origination.
SPEAKER_01Indeed. Yes, I mean that's been the calling card, that's been the the kind of uh KPI of value that we've been tracking with all of our clients to date. It's been on net new opportunities, you know, finding relationships in the market that they they weren't aware of. It's then monitoring the known universe of opportunities to maintain that regular cadence of outreach for relationship management, and specifically to know when will be the moment they're most likely to transact. So that's the kind of operational calling card of our proposition. But the strategic point is, as you say, having assimilated this powerful internal market database, it is then able to feed the whole of their AI roadmap that they want to achieve in other workflow throughout the firm. And it's that second strategic point, with us being that foundation stone of data for their AI roadmap, which is capturing the attention of the managing partners and the board level. And I would I would say, you know, we're well aware that there's huge competition for fundraising. The LPs are asking very pointed questions to the uh the managing partner level of you know, what's your competitive edge over the next decade? How are you leveraging data? How are you leveraging AI? And those firms that don't have a good answer for that, it's it's having a material effect on their ability to fundraise. Now, as I said earlier, our our raison d'etre is to help P firms to build that architecture faster and with low lower risk in order to achieve those outcomes in a quarter, as opposed to an 18-month self-build time.
SPEAKER_00Yeah, I um well timing is is part of this, and and just going back to where right at the beginning of the ROI piece is that if that sort of lead time is long, how how do you when do you know that you're that you're getting ROI from this sort of, I suppose, rewiring of their deal origination process?
SPEAKER_01Yeah, so it's a big part of the development we've made in the last six months is tracking those KPIs across all the value props, which is enabling our sponsor, whether it's the tech team, the head of origination, the deal team, to present the ROI to the board. So just a few a few stats from a client I was chatting to this morning. Um, they've had 147 new relationships on their radar in the last two and a half months. So those are 147 good fit opportunities. They have been recommended by the system while the deal makers or the origination team were head down on a deal. So that's been passively recommended. Obviously, there's been a human review. Um, our our proposition is not to pick the winners and to replace the skill of the originator or the investment professional, but it's to find the best fit. So it's a resource allocation. So you're now looking at, let's say, 280 in their case, 280 recommendations, which you whistle down to the 147 that will be passed into their CRM and now tracked every week on an ongoing basis to know you know how do we outreach, what is um, what is a reason to get in touch and have an idea on when they'll be ready to transact.
SPEAKER_00I'm an advocate for the human in the loop here, but it feels like that sort of monitoring and and touch points could be automated as well.
SPEAKER_01Not in the next few years, in my opinion, because a key part of deal making is building that relationship. So another another stat, another firm has interacted with 302 actionable insights that have been pushed to them by their filament Sipta system with very tailored news, financial metrics, new hires, uh context on that opportunity, which gives them a reason to outreach and build that relationship. So it's about empowering the deal maker, giving them the Iron Man suit so they can do their job better and more effectively. And just to give you a little nugget of what Gen AI is bringing to the table there, having assimilated that actionable insight, translating that into an outreach email, which the deal maker can request, and they get a draft email that they can send to the founder or chairman, and typically they can then top and tail it with a personal touch and and then send it off, and that whole workflow can be conducted in a couple of minutes, and so it's just facilitating the way they build relationships for deal making.
SPEAKER_00Yeah, um, and they can do a hell of a lot more than they were able to previously. Would you say that then well yeah, just on that, then those 147, would in the traditional sense would that have been halved because it's just they just wouldn't be able to cope at that level? Is something changed there in terms of deal volume?
SPEAKER_01Yeah, two fronts. Even if you've got a dedicated origination team, they've only got a limited capacity. And when funds are doing well and they've got resources to deploy, assets under management, you know, they've got a new fund. The the option to the head of origination is do I get you know more, more, more, grow my team to be able to cover the market better. There is just the the manual task of going to the different uh SaaS sourcing platforms, all of which are fantastic in their different ways. But to get that full picture of an opportunity, you're going to the website, you're going to three or four different SaaS platforms, you're going to LinkedIn, you're doing a Google search, you're checking the media. Having that assimilate into one pane of glass is a huge time saving. And also turning origination from a batch job that you do periodically and you build up a list in Excel, or you build up a PowerPoint for a market mapping exercise into a list which is now saved and tracked internally forevermore, is just a step change in the whole efficiency piece of deal making and originating. So that is the step change that a proprietary system gives you if you move your addressable market into an internal tracked database.
SPEAKER_00One of the things that um well look my uh network is is still relatively new. And so you'd expect those um people in the network right now are those that are most curious or those that have been sort of assigned to to looking at new opportunities with with AI. But it's clear, you know, that there are cultural hurdles to get past to bring about proper adoption.
SPEAKER_01Yeah, I mean it's been it's been building up groundswell over the last few years. I'd say our our early clients were certainly the the the early movers back in sort of 2019 to 2022, but in the last couple of years, this is going mainstream. I mean, almost every PE I speak to has at least got an in-house pet project where they're exploring data integration, codifying their scoring, building um gen AI workflows or experimenting with them at the very least. In order to operationalize it, it it would be very sensible to get a partner on that journey, a tech partner that has done it before, that can accelerate, can de-risk, and and reduce the cost dramatically of trying to assemble a team to build in-house. And even some of the clients that started building in-house a few years ago are now coming back to us, maintaining the name and the branding of their platform, but now white labeling SIFT are under the hood so that they can take it to the next level. Uh, because these PE firms are brilliant at what they do, they're not software houses. And so the make or buy decision is a little nuanced here because we're not talking about buying another SaaS platform, another five seats on another SaaS platform. It is it is about building in-house, but it's building with a partner like Philip and Sifter to ensure you've got um quicker time to value, you're responsive to the market, you've got an innovation partner making sure you're not just keeping up, but actually ahead of the pack.
SPEAKER_00Is there a um is there some things that they could be doing? You know, you've talked about the pet projects, but you know, before they would reach out to Filament Sift though, what should they be doing themselves right now? What's really important, those first. Few steps.
SPEAKER_01Yeah, I think identifying the owner within the organization that is, if you will, the product manager for their data and AI strategy. And we've seen that being a side project from either the value creation team, the tech team, or the deal team. And all three of those have their pros and cons. But as we all know, a good product manager is about setting the vision, getting the organization behind it, um, procuring well, um, connecting your internal systems of record, which is what we're talking about here essentially, it's bringing your market intelligence in-house, mixing it with the proprietary intelligence you have in your CRM or in your SharePoint or in your port co-management system, and creating that organizational brain. So, as a first step, whether it's a net new hire, which many firms are now doing with like heads of data, heads of AI, their first data science recruit, or just having that kind of focal point, the owner of that roadmap who reports to the managing partners. And and you know, this is seen by the managing partners as a strategic imperative. I I think identifying that person is a key first step.
SPEAKER_00Yeah, yeah. That that makes a lot of sense. Well, look, you've taken a you know, it's taken some time to get here, and there's been advances in technology along the way to support your your mission. But what's what's next? I mean, we touched we touched on it maybe earlier, but but in your view, um not looking for you to reveal with the secrets of your roadmap, but you know, what where do you see it going?
SPEAKER_01Yeah, well, I think every P firm, every corporate finance firm working in private markets needs to be assimilating their data and and connecting it and creating the fuel for the the organizational brain. And that and that technology exists, and all firms will need to have that form of architecture, knitting together their systems of record and supercharging their deal makers. And if you if you think about the raison d'etre for a private equity firm, it's finding and closing the best deals, and then it's adding value through the life of that investment. And those that do that the best in a in a data and AI empowered way will get the best returns and be the most successful in the market. It's as simple as that. I think the Gen AI tool set is is super interesting, but that foundational piece, having that accumulated data set that you you own and you enrich every day with the combined intellectual property of your team, that is the foundational step. Once you have it, there's some very cool work we're doing with some of our most forward-thinking clients on then applying different ways of working, different ways of interacting with your in-house intelligence. Um, and just to give you a feel for that, there's a piece about enriching your team's experience where they work today. So if they have a favorite SaaS platform that is the one they've used all their career and is their go-to, then don't fight against that. Don't try and port them onto another platform. Actually, seed the intelligence where they work today. So they could be looking at an opportunity on the website or in their favorite sourcing tool, SaaS sourcing tool, and a pop-up of the company's brain, which says, well, the last colleague to meet them was X on this date. Here are the notes from the founder, here is the in-house codified scoring on how we rate them as an opportunity for our different funds. And here is some Gen AI workflow that we can initiate. For example, I want a one-pager on this on this opportunity combining our internal and external data in the format that we like to send to IC. Here is a ChatGPT-like capability. I want to ask questions of this company with respect to our proprietary database. So just to be clear, that's different to ChatGPT asking the web and getting the web's opinion. This is the in-house brain that you're querying, and therefore it is enriched with information only you know as a firm. And that is that is a game changer. You're looking within your walled garden for information about that opportunity and how we as a firm rank it against its competitors or see the investment opportunity.
SPEAKER_00Do you think there's more of a catalyst when there's a new raise going on to adapt or rework the processes? Is there a certain timing that would make sense to shift to you know to something like this?
SPEAKER_01Yeah, so from our experience, um, clients have um got going either pre-fundraise or post-fundraise. The pre-fundraise gives them the ammunitions to go to the LP and say, look, this is the platform we're building. We're actually working with a few brand new funds who have started with this data infrastructure with CIFTRA at their core from the get-go, and it's enabled them to go fundraising with a data and AI first approach. But for for other more established funds, it it's often useful pre-fundraising to be able to show the LPs here's our in-house engine. They they name it something fancy or they name it GP name AI, and it's undoubtedly making a material difference to their ability to fundraise by showing the LPs they're ahead of the game.
SPEAKER_00Well, I think that's uh that's a good point to to leave it, Phil. Um, really fascinating. I look into this particular area. Um, you know, really good um sharp insight into origination and how that's how that's uh evolving. And you know, I'm sure if they're not already, you know, firms will be rethinking what priority means after hearing this. So thank you so much for your time today, Phil.
SPEAKER_01Thanks, Steve. Thanks for having me.
SPEAKER_00Um I'll place uh links to Phil and Funament Sister in the show notes. I'm sure you can run a search for him and uh Fulament as well. It just leaves me to just do a little plug um for our next session in September, a full day AI strategy day. We'll be looking at you know immediate practical applications of AI throughout the deal lifecycle, supported by keynotes, case studies, workshops, ensuring that you gain not just ideas, but things you can action immediately. So check out aipathfinder.co for more information. But that's it for now. So see you soon.